Copy article

Amazon Ring Rolls Out AI Facial Recognition to US Doorbells Despite Privacy Concerns

ended 11. December 2025

After an announcement in September, Amazon has launched its "Familiar Faces" AI-powered facial recognition feature for Ring video doorbells across the United States, enabling users to create catalogues of up to 50 faces for personalised notifications. The feature, now rolling out, allows device owners to label regular visitors including family, friends, delivery drivers, and household staff.

The system processes biometric data in the cloud and sends customized alerts such as "Mom at Front Door" instead of generic person detection notifications. Amazon states the feature is opt-in, with face data encrypted and unnamed faces automatically deleted after 30 days. The company claims it does not use this data for AI model training.

However, the rollout has attracted immediate criticism from privacy advocates and lawmakers. US Senator Ed Markey (D-Mass.) has called for Amazon to abandon the feature, while the Electronic Frontier Foundation (EFF) has raised concerns about consent and data protection. The feature cannot launch in Illinois, Texas, and Portland, Oregon due to existing biometric privacy laws.

Amazon's Ring has a documented history of privacy failures. In 2023, the company paid a $5.8 million FTC fine after employees and contractors were found to have unrestricted access to customer videos. The company previously partnered with law enforcement agencies, giving police direct access to request doorbell footage, and recently partnered with Flock Safety, an AI surveillance provider working with ICE and federal agencies.

We'd like your views:

  • Should individuals passing by Ring cameras have the right to know they're being enrolled in facial recognition databases?
  • How do biometric privacy protections need to evolve when surveillance technology moves into residential settings?
  • What safeguards should exist before facial recognition data can be shared with law enforcement?
  • Should delivery workers and service providers have legal protection from residential biometric scanning?
  • How can consent frameworks address situations where the person being scanned has no relationship with the device owner?
  • What enforcement mechanisms are needed when companies with poor security track records deploy biometric technologies?

4 responses from the Newspage community

Copy all

Star Quote
Copy

Amazon's been so busy asking "Can this tech recognise faces?" that it forgot to ask "Should everyone walking past your door be conscripted into your personal biometric database without consent?"

Ring's rolling this out while Meta's AI glasses are already letting US strangers identify people on the street, and UK police are testing AI-powered van cameras that scan number plates and faces simultaneously. This isn't three separate privacy concerns. It's the infrastructure for pervasive surveillance being built one "convenient feature" at a time, by companies with documented security failures and cosy relationships with law enforcement.

We're normalising a world where you can't walk down a street, approach a door, or drive past a van without being catalogued, analysed, and stored by systems you never agreed to join. Everyone else's anonymity is at risk, watched by a system that's increasingly difficult to dismantle.
Copy

Amazon’s new “Familiar Faces” feature is clever tech. Who wouldn’t want to know exactly who’s at the door? The issue isn’t AI. It’s the company behind it. AI is like a high-performance camera: powerful, useful, reassuring. But once the footage leaves your house, you’re no longer the one holding the lens.

Amazon has a history of saying data will be used “to help you”, then quietly finding new uses you never agreed to. Remember when Google taught creators how to rank, only to scrape that same content years later to train its own AI? That’s the playbook.

So yes, the technology is exciting. But trusting Amazon with biometric data is like giving your house keys to someone who’s already copied them once. The risk isn’t what the tool does today. It’s what the company might do with your face tomorrow.
Copy

Your doorbell should not get more say over a worker’s face than the worker does. Delivery staff and tradespeople absolutely need legal protection from residential biometric scanning. This is workplace tech by the back door. They are doing their job, often on tight targets, with zero real freedom to say “no thanks, please don’t store my face”. That is not consent, it is coercion in a hi-vis jacket.
Facial recognition is effectively a portable ID card stapled to someone’s head. Once it exists, it can be matched, shared, misused. So before any of this touches law enforcement, the default should be strict: clear warrants, strict time limits, proper audit trails, and workers’ rights built in. If a company already has a dodgy privacy record, the bar should be higher, not “oh well, try again”. Fines that bite, orders to delete data, and a very real risk of losing the right to process biometrics at all.
Copy

The "sci-fi" dystopia isn't coming; it’s already here, sold to us as a doorbell feature. By pushing facial recognition to the cloud rather than keeping it local, Ring is creating a honeypot for hackers.

We have to look at the downstream risks. We already have AI tools capable of generating realistic video from static images. If a bad actor accesses Ring's "Familiar Faces" database, they don't just get a photo, they get the keys to your digital identity. They can generate videos of you doing things you never did, or access systems locked by your biometrics.

It is baffling that we are trusting a company with a $5.8m fine for privacy violations to hold this data. They hold your history hostage behind a subscription wall, and now they want your biometrics too? This isn't innovation; it’s a security vulnerability masquerading as a premium feature. Keep your data local, or don't collect it at all.